As previously mentioned, I am having good luck working with XQuery. Today, I was experimenting with pulling multiple RSS feeds together into a single feed. Here is a sample XQuery and result:

the XQuery looks like (note: URLs shortened for appearance):

<rss version="0.92">
{
let $url := (
	"http://icite.net/blog/. . .,
	"http://jena.hpl.hp.com:3030/blojsom-devt/blog/. . .,
	"http://www.blojsom.com/blog/. . . 
	)
let $a := for $i in $url 
	let $b := doc($i)/rss/channel 
	return
	$b/title union $b/item

let $a := <t> {$a} </t>	
return 
	<channel>
	<title>blog mosh</title>
	<description>feeds from: {
		fn:string-join($a/title/text(), ", ")
		} moshed up via XQuery
	</description>
	<language>en</language>
	{ $a/item }
	</channel>
}
</rss>

(Unedited, source XQuery here.)

Basically, the XQuery says: 1) here are some URLs, 2) grab the title and item nodes within the /rss/channel nodes found at those URLs, 3) return stuff in a format that matches the RSS 0.92 format. (There is a little trick in there in that the new feed title lists the titles of each source feed.)

Because there are three source URLs, the result is a mashup RSS feed with three entries, one each from my blog, the HP Semantic Blogging Demonstrator, and David Czarnecki's blog respectively. Here is the full query result XML file.

You might notice that I chose three blogs that all run on blojsom. I did this because I know how each blojsom blog entry can be seen, by itself, as RSS. This example works with single posts in RSS only to keep the result brief—full feed mashups also work!

For my next experiment, I am thinking of creating an interactive XQuery form on this site (i.e., a web UI to the XQuery interface of Saxon). I also want to experiment more with combining X/HTML pages, but I think I need to add a Tidy process in the loop to get XML clean enough for Saxon / XQuery to process.